After a very long break, we decided to experiment with making in-house PCBs.
Normally we design PCBs for production, and get protos from places like
4pcb.com.
Last night we set up a makeshift lab, but ran into a small hiccup. This
morning, after a trip to Fry's Electronics to get the correct type of
developer, we developed & etched our first PCBs:
Last time I used a presensitized PCB method was all the way back in 2001,
and I've completely forgotten how quick and painless it is. We will
definitely use this process for all of our in-house projects going forward.
I think the part that made it sound less attractive before, was the thought
of having to do all the drilling and the associated noise/dust/extra time.
> After a very long break, we decided to experiment with making in-house PCBs.
> Normally we design PCBs for production, and get protos from places like
> 4pcb.com.
>
> Last night we set up a makeshift lab, but ran into a small hiccup. This
> morning, after a trip to Fry's Electronics to get the correct type of
> developer, we developed& etched our first PCBs:
>
> http://maksimov.org/piclist/pcb_etching/pcb_etching.jpg
>
> I guess there is something magical about the first time you see the bare
> copper disappear before your eyes. :)
>
> Finished "product":
>
> http://maksimov.org/piclist/pcb_etching/overcurrent_alarm.jpg
>
> Last time I used a presensitized PCB method was all the way back in 2001,
> and I've completely forgotten how quick and painless it is. We will
> definitely use this process for all of our in-house projects going forward.
>
> I think the part that made it sound less attractive before, was the thought
> of having to do all the drilling and the associated noise/dust/extra time.
>
> Just thought I'd share. :)
>
> Best regards,
>
> Vitaliy Maksimov
> ScanTool.net, LLC
> +1 623 582-2366
> http://www.scantool.net
>
Would you care to provide a list of materials? Perhaps with methods?
I'm also looking at doing this again. One distinct advantage of surface
mount parts - no need for holes.
Although, if I had done your design, I would have put pads down anyway
in the corners to mark drill locations in the event I wanted to mount
the board with standoffs later. And perhaps run the four I/O lines to a
1/10" header pattern in case that looked useful in the future. If
un-needed, don't drill 'em. But the effort to put them in the artwork
is minimal. Just another view...
Thanks. I'm having to look at how I do my one-off boards as
Radio Spares have stopped stocking the etch-resist transfers.
I've had a go with toner-transfer and that's worked OK, but a
bit too hit-and-miss when it comes to SMT. I could still do SMT
boards by hand, although I think a photo-resist system might be
worth investing in, especially for repeat boards. Hopefully I'll
expand into manufacture this year, most likely bespoke small runs
which are awkward re: time vs money. If resources permit, or sales
demand, then I'd get batches of boards made professionally but
otherwise I need a reliable DIY method
Vitaliy wrote:
> After a very long break, we decided to experiment with making in-house PCBs.
> Normally we design PCBs for production, and get protos from places like
> 4pcb.com.
... "only problem is, it takes too frickin' long!"
Never thought I'd find a suitable time to use a Die Hard 3 reference on
the PICLIST...
> Last night we set up a makeshift lab, but ran into a small hiccup. This
> morning, after a trip to Fry's Electronics to get the correct type of
> developer, we developed & etched our first PCBs:
>
> http://maksimov.org/piclist/pcb_etching/pcb_etching.jpg
Hm. Is that ammonium persulphate etchant? Not a fan of ferric chloride
then, I guess?
> I guess there is something magical about the first time you see the bare
> copper disappear before your eyes. :)
Nice board. Did you immersion-tin it as well, or is that just the way
the light's reflecting off the copper?
> Last time I used a presensitized PCB method was all the way back in 2001,
> and I've completely forgotten how quick and painless it is. We will
> definitely use this process for all of our in-house projects going forward.
Yeah, it certainly beats toner-transfer. Cheaper, possibly a little
slower, but definitely better. Hey, two outta' three ain't bad.
> I think the part that made it sound less attractive before, was the thought
> of having to do all the drilling and the associated noise/dust/extra time.
Which is why you go SMD. "Holes, what holes? For components? Why would I
need those?"
Marcel Duchamp wrote:
> Would you care to provide a list of materials?
Sure, I don't have time right now but I'll put together a list of p/ns
later.
>Perhaps with methods?
Not sure what you mean. Like, "tips and tricks"? :)
> I'm also looking at doing this again. One distinct advantage of surface
> mount parts - no need for holes.
>
> Although, if I had done your design, I would have put pads down anyway
> in the corners to mark drill locations in the event I wanted to mount
> the board with standoffs later. And perhaps run the four I/O lines to a
> 1/10" header pattern in case that looked useful in the future. If
> un-needed, don't drill 'em. But the effort to put them in the artwork
> is minimal. Just another view...
I did not design the PCB, and the explicit goal was to get it done in the
shortest amount of time (the main goal was to see how feasible this
PCB-making process was). The application virtually insures that standoffs
and the headers will never be used. Plus, in general we follow the YAGNI
philosophy -- much time/effort gets wasted on things that are never used,
even when they appear to be "free".
Philip Pemberton wrote:
>> After a very long break, we decided to experiment with making in-house
>> PCBs.
>> Normally we design PCBs for production, and get protos from places like
>> 4pcb.com.
>
> ... "only problem is, it takes too frickin' long!"
Exactly! :) Plus, what's the fun in that?
It can also be more cost effective, as long as you don't have four people
just standing there watching the fishtank for 15 minutes. ;)))
>> Last night we set up a makeshift lab, but ran into a small hiccup. This
>> morning, after a trip to Fry's Electronics to get the correct type of
>> developer, we developed & etched our first PCBs:
>>
>> http://maksimov.org/piclist/pcb_etching/pcb_etching.jpg
>
> Hm. Is that ammonium persulphate etchant? Not a fan of ferric chloride
> then, I guess?
Used probably more than a bathtubful of that nasty brown stuff back in the
day.. Ammonium persulphate is so much nicer: you can see the PCB, it doesn't
stain hands, clothes, or furniture, doesn't require the use of CLR after you
dump it in the sink..
That's the green resist. The mfr said you can leave it on and solder through
it, so we did.
>> I think the part that made it sound less attractive before, was the
>> thought
>> of having to do all the drilling and the associated noise/dust/extra
>> time.
>
> Which is why you go SMD. "Holes, what holes? For components? Why would I
> need those?"
On 2/26/2010 3:22 PM, Vitaliy wrote:
> Marcel Duchamp wrote:
>> Would you care to provide a list of materials?
>
> Sure, I don't have time right now but I'll put together a list of p/ns
> later.
Perfect.
>
>> Perhaps with methods?
>
> Not sure what you mean. Like, "tips and tricks"? :)
Right; just things you learned that could help others avoid various and
sundry pitfalls. Plus, you had to choose from things like toner
transfer or sensitized boards, etc. If you made choices, which ones did
you choose and how did it work out. What if anything would you do
differently...
> It can also be more cost effective, as long as you don't have four people
> just standing there watching the fishtank for 15 minutes. ;)))
Let's see... 15 minutes times four people is an hour... still only £10
($15 or so) assuming that's their hourly rate (probably a bit low, but
still not worth worrying about).
Now, all those rubbernecking staff walking past.... them's the folk you
need to be watchin' out for :)
>> Hm. Is that ammonium persulphate etchant? Not a fan of ferric chloride
>> then, I guess?
>
> Used probably more than a bathtubful of that nasty brown stuff back in the
> day.. Ammonium persulphate is so much nicer: you can see the PCB, it doesn't
> stain hands, clothes, or furniture, doesn't require the use of CLR after you
> dump it in the sink..
You probably shouldn't dump it in the sink....
I was always told the best way to dispose of FeCl3 was to mix it with
sodium carbonate (or alternatively sodium bicarbonate / baking soda or
sodium hydroxide PCB developer), then keep adding Na2CO3 until the pH
passes neutral and starts to go alkaline. Run it through a sheet of
filter paper -- the liquid is water with dissolved sodium salts, the
sludge is a mix of iron and copper and should go to a hazchem dump.
Needless to say, don't do this in your FeCl storage bottle.
Apparently the easiest etchant to work with is cupric chloride -- to
start it off, you use a mix of hydrogen peroxide and hydrochloric acid,
then etch a ton of copper with it. Dump a load of scrap copper in the
tank and bubble air through it. After a while your HCl+H2O2 etchant
becomes CuCl etchant. There's a chapter on this in "Electronic Prototype
Construction" by Steve Kasten (Blacksburg / Sams, published some time in
the late 70s). I think someone's typed in said chapter and published it
online...
>> Nice board. Did you immersion-tin it as well, or is that just the way
>> the light's reflecting off the copper?
>
> That's the green resist. The mfr said you can leave it on and solder through
> it, so we did.
I tried that with some of mine -- they made a really nasty fishy smell
while I was soldering them, so I got a bottle of acetone and gave it a
dip. Followed that with a fairly liberal spray-coating of Electrolube
SMFL solder flux.
>> Which is why you go SMD. "Holes, what holes? For components? Why would I
>> need those?"
>
> :)
>
On a more serious note, I'm doing a PCB design for a piece of fairly
low-volume DAQ hardware. >95% of the holes in the board are vias, and
almost all of said vias are there either for thermal relief or to keep
the impedance of the power planes down. The SMPSU stuff uses big, meaty
power and ground planes -- GND on the bottom of the board. To pull GND
to the top, I create a second, smaller GND plane on the top layer and
link it with a few vias arranged in a grid.
Well, it's what TI suggested in the reference design... "To keep ground
plane impedance and EMI at manageable levels, and to improve thermal
performance, use multiple vias to link to the ground plane."
I don't know if this is bullcrap, but it doesn't really do any harm.
Plus the six thermal vias underneath the QFN power management chip might
make the thermal pad a bit easier to solder... unless they draw all the
heat from my hot-air station down onto the ground plane (which would be
a Bad Thing).
Marcel Duchamp wrote:
>>> Would you care to provide a list of materials?
MG Chemicals
CAT #603
Single Sided Presensitized Copper Clad Board
Developer (Cat # 418)
For use with 600 series Presensitized Boards
Ammonium Persulphate
I think we bought the stuff from Circuit Specialists, but I found out that
you can get the same stuff for about the same price from Fry's Electronics.
>>> Perhaps with methods?
>>
>> Not sure what you mean. Like, "tips and tricks"? :)
>
> Right; just things you learned that could help others avoid various and
> sundry pitfalls. Plus, you had to choose from things like toner
> transfer or sensitized boards, etc. If you made choices, which ones did
> you choose and how did it work out. What if anything would you do
> differently...
Ok, let's see...
The first problem was with the transparency, we printed to a black-and-white
Brother MFC8860 printer and the black was sort of spotty (you could really
see it in the big pads). In retrospect, it would probably still work fine,
but we decided to play it safe and reprinted the PCB on a color printer
(Dell 3110cn) which has much better black coverage.
One big mistake I made, was buying the "Negative Developer" which doesn't
work with presensitized boards. It was like water, in fact the next morning
when I got the right developer I pulled the PCB out of the trash bin for
testing, and it came out almost perfect. Since it was "just a test" , and
didn't expect it to come out so well, I didn't have a rinse tray nearby and
over-developed it).
We bought a photo frame, which sounded good in theory but in practice is too
cumbersome. For the second batch we simply put the PCB on a book and put the
glass on top of it. I think when we were fiddling with the photo frame, we
scratched the transparency and did not notice the problem until after we
etched the boards.
Exposure doesn't seem to be a very critical process. We used the "150W
equivalent" CFL and also tried a real 150W incandescent for about 5 minutes
each, results were similar. I'm partial to the CFL, theoretically it's a
better light source (more blue light), it doesn't get as hot, and it is more
uniform.
The actual development process is very fast, it literally takes less than a
minute. We didn't use a brush as suggested in the instruction sheet, just
sloshed the developer back and forth in the tray. Once the exposed resist is
mostly gone, dunk the board in a rinse tray. Err on the side of
underdeveloping (you can always put the board back in the developer).
Etching was easy, preheat the solution to 55C, it takes about 10 or 15
minutes per board. I love Ammonium Persulphate, and the clear fishtank (my
old one was green). Remember that etching is a slow process until you start
seeing the copper disapper. Then it's lightning fast, and you need to pull
it out just after all the exposed copper is gone to avoid over-etching.
>On 2/26/2010 2:23 PM, Vitaliy wrote:
>> After a very long break, we decided to experiment with making in-house PCBs.
>> Normally we design PCBs for production, and get protos from places like
>> 4pcb.com.
>>
>> Last night we set up a makeshift lab, but ran into a small hiccup. This
>> morning, after a trip to Fry's Electronics to get the correct type of
>> developer, we developed& etched our first PCBs:
>>
>> maksimov.org/piclist/pcb_etching/pcb_etching.jpg
>>
>> I guess there is something magical about the first time you see the bare
>> copper disappear before your eyes. :)
>>
>> Finished "product":
>>
>> maksimov.org/piclist/pcb_etching/overcurrent_alarm.jpg
>>
>> Last time I used a presensitized PCB method was all the way back in 2001,
>> and I've completely forgotten how quick and painless it is. We will
>> definitely use this process for all of our in-house projects going forward.
>>
>> I think the part that made it sound less attractive before, was the thought
>> of having to do all the drilling and the associated noise/dust/extra time.
>>
>> Just thought I'd share. :)
>>
>> Best regards,
>>
>> Vitaliy Maksimov
>> ScanTool.net, LLC
>> +1 623 582-2366
>> http://www.scantool.net
>>
>
>Would you care to provide a list of materials? Perhaps with methods?
>I'm also looking at doing this again. One distinct advantage of surface
>mount parts - no need for holes.
>
>Although, if I had done your design, I would have put pads down anyway
>in the corners to mark drill locations in the event I wanted to mount
>the board with standoffs later. And perhaps run the four I/O lines to a
>1/10" header pattern in case that looked useful in the future. If
>un-needed, don't drill 'em. But the effort to put them in the artwork
>is minimal. Just another view...
>
>Mike Harrison wrote:
>> check out my page http://www.electricstuff.co.uk/pcbs.html for lots of info on
>> making good PCBs.
>
>Why do you use UV? What is the advantage over "normal" light sensitive PCBs?
>
All 'normal' PCB material I've ever encountered is primarily sensitive to UV, to the extent that you
can handle it in normal light for the lengths of time necessary for PCB making.
I've never actually tried exposing with visible leight but I suspect exposure times would be pretty
long.
> After a very long break, we decided to experiment with making in-house PCBs.
> Normally we design PCBs for production, and get protos from places like
> 4pcb.com.
>
> Last night we set up a makeshift lab, but ran into a small hiccup. This
> morning, after a trip to Fry's Electronics to get the correct type of
> developer, we developed & etched our first PCBs:
>
> http://maksimov.org/piclist/pcb_etching/pcb_etching.jpg
>
> I guess there is something magical about the first time you see the bare
> copper disappear before your eyes. :)
>
> Finished "product":
>
> http://maksimov.org/piclist/pcb_etching/overcurrent_alarm.jpg
>
> Last time I used a presensitized PCB method was all the way back in 2001,
> and I've completely forgotten how quick and painless it is. We will
> definitely use this process for all of our in-house projects going forward.
>
> I think the part that made it sound less attractive before, was the thought
> of having to do all the drilling and the associated noise/dust/extra time.
>
> Just thought I'd share. :)
>
> Best regards,
>
> Vitaliy Maksimov
> ScanTool.net, LLC
> +1 623 582-2366
> http://www.scantool.net
>
M. Adam Davis wrote:
> I was just thinking about starting doing this again myself, since all
> of my stuff is surface mount I can skip most of the drilling.
>
> I expect for the more complex boards I can use 1206 0 ohm resistors as
> jumpers (ie, to be the second side).
>
> But it would probably beat the occasional need to do a dead bug
> prototype of a fine pitch part!
Dead-bugging still has it's place.
I've been known to use "ugly construction" or variants of "Manhattan"
construction for quick prototypes. Most notably a thermocouple amp I
built up to test a Maxim thermocouple controller chip (MAX6675? the one
with the built in cold-junction compensator, A/D and SPI interface).
Good fun for making little modules to hook up on a breadboard, but
utterly hopeless for anything remotely complex...
Mike Harrison wrote:
>>Why do you use UV? What is the advantage over "normal" light sensitive
>>PCBs?
>>
>
> All 'normal' PCB material I've ever encountered is primarily sensitive to
> UV, to the extent that you
> can handle it in normal light for the lengths of time necessary for PCB
> making.
>
> I've never actually tried exposing with visible leight but I suspect
> exposure times would be pretty
> long.
The boards we used are sensitive to normal light. You can handle them either
in subdued normal light or yellow light. Exposure time is ~5 minutes, using
either incandescent or fluorescent bulbs.
I'm surprised nobody's mentioned a CNC mill in regards prototyping yet. I
would've thought it would be quick and repeatable (especially if lots of
drilling is required), although there is the initial outlay for a maching - but
if you're doing a lot of prototypes this probably won't be much of a concern.
Plus you can use it to cut holes in boxes, etc. and fabricate a whole host of
other hardware-related bits and bobs.
Has anybody got any practical experience of CNC-ing prototype boards ? This
is something I'm interested in - eventually I'd like to build my own CNC rig
but it's just time and money...
On Sun, Feb 28, 2010 at 1:07 PM, Peter Restall <EraseMEpetespam_OUTTakeThisOuTrestall.net> wrote:
>
> I'm surprised nobody's mentioned a CNC mill in regards prototyping yet. Â I
> would've thought it would be quick and repeatable (especially if lots of
> drilling is required), although there is the initial outlay for a maching - but
> if you're doing a lot of prototypes this probably won't be much of a concern.
> Plus you can use it to cut holes in boxes, etc. and fabricate a whole host of
> other hardware-related bits and bobs.
>
> Has anybody got any practical experience of CNC-ing prototype boards ? Â This
> is something I'm interested in - eventually I'd like to build my own CNC rig
> but it's just time and money...
>
> Regards,
>
> Pete Restall
I had access to a machine several years ago in college. I was never
able to get very good boards with it. Milling the PCB results in
grooves on the surface which makes a rough and unsightly board. There
is no solder mask or silkscreen. Unless the bit is brand new it will
do a poor job milling the copper, leaving burrs that will short out
traces. IIRC the minimum trace/space with this machine was 10mil/10mil
(0.01 inch) which precludes any modern parts. You can do two layer
boards but you can't do plated through vias or holes.
All-in-all it's a bum deal for a machine that is (at minimum) several
thousand dollars. It hasn't really caught on because people who have
the capital to pay for it are probably willing to pay to have boards
expedited, and are probably also doing "state-of-1970" circuit boards
that are too advanced for a milling machine to produce.
> I'm surprised nobody's mentioned a CNC mill in regards prototyping yet. I
> would've thought it would be quick and repeatable (especially if lots of
> drilling is required), although there is the initial outlay for a maching - but
> if you're doing a lot of prototypes this probably won't be much of a concern.
> Plus you can use it to cut holes in boxes, etc. and fabricate a whole host of
> other hardware-related bits and bobs.
>
> Has anybody got any practical experience of CNC-ing prototype boards ? This
> is something I'm interested in - eventually I'd like to build my own CNC rig
> but it's just time and money...
>
> Regards,
>
> Pete Restall
At my secondary school they used to produce simple PCBs (one sided, no
silkscreen, no mask) with very short turnaround. The results were
adequate for the simple circuits that were created.
I built a CNC machine from a milling machine using EMC2 software and an
interface of my own design. While I did this orginially to do aluminum
with, I recently have been doing PCBs with it. For one board that uses a
2mm x 2mm chip with 3 pins on opposite sides I use a 20 degree engraving bit
for a gap of about .018". For less critical work I use a .029" end mill set
only deep enough to remove the copper and not the board. You can see a
board I am currently selling at http://www.dv-fansler.com/dvf%20Technologies/lx200_keypad_sim.htm along with
3 pictures. This board was done with the .029" end mill. I have done over
20 boards on the same end mill and they still look like the first one.
BTW - I do my layout in AutoCAD and then convert it to ngc with a free lisp
program called Realize.
M.L. wrote:
> I had access to a machine several years ago in college. I was never
> able to get very good boards with it. Milling the PCB results in
> grooves on the surface which makes a rough and unsightly board. There
> is no solder mask or silkscreen. Unless the bit is brand new it will
> do a poor job milling the copper, leaving burrs that will short out
> traces. IIRC the minimum trace/space with this machine was 10mil/10mil
> (0.01 inch) which precludes any modern parts. You can do two layer
> boards but you can't do plated through vias or holes.
>
> All-in-all it's a bum deal for a machine that is (at minimum) several
> thousand dollars. It hasn't really caught on because people who have
> the capital to pay for it are probably willing to pay to have boards
> expedited, and are probably also doing "state-of-1970" circuit boards
> that are too advanced for a milling machine to produce.
We looked at CNC for prototype production, but as you say it did not look
like it was going to be cost effective. It would take years for the machine
to pay for itself (compared to getting the protos from a board house), and
you don't get a "production quality" PCB.
> I bought a used LPKF from eBay a couple years ago for hobbyist
> usage. Less than half the price of the cheapest new unit...
>
> I don't know if it was worth it. I'm pretty sure I'm still far
> behind where I'd be $-wise if I'd sent every board I've made
> (including the N iterations of some of them) to a professional
> fabricator. It's nice to be able to go from eagle design to
> SS PCB in a couple hours, especially for the tiny boards that
> aren't quite right that I tend to make, but often the several
> hours of baby-sitting needed don't come along for quite a while,
> and I could've sent them out.
>
> It's a pain not having Plated through holes. Yeah, the newer
> systems have some PTH system of somesort, but it adds to the
> time and complexity of fabrication...
>
> It's a pain not having a soldermask.
>
> Tools and other expendable items are expensive; you can buy surplus
> partially used drills for cheap, but the mechanical etching bits are
> uncommon, without many providers, and rather expensive (and they
> wear out pretty quick.)
>
> I have my doubts about it working very well below about 1mm pin
> pitch. While the specs say you can get 0.2mm isolation (with the
> standard tool), you're looking at a V-shaped trench, so if you
> were hoping to do .5mm TQFP with 0.3mm lands 0.2mm apart, you're
> at the very edge of possible, and you get a pattern in the third
> dimension that is exactly wrong for helping to position the packages.
> (trenches for the leads to fall in, between the pads.) Did I mention
> that it's a pain not having soldermask?
>
> Steps for more than 2 layers would add additional complexity and
> cost. Did I mention that it's a pain not having PTH ?
>
> It's really noisy. I think they rig the trade-show demos where it
> seems
> to run at only "moderate" noise levels. Or maybe the vacuums have
> gotten
> better since the one I have.
>
> On the bright side, it's a pretty good CNC drill even if everything
> else stops working. You can easily have complex board outlines and
> internal cutouts and such that are rare to find in "prototype" PCB
> houses. You can do signs and (2D) CNC-router type projects other
> than PCBs. The software works ok, isn't hard to use, and seems to
> offer capabilities far beyond "free" stuff (like EAGLE's OUTLINES.ULP)
Thanks for the input; some interesting points there. I've only managed to
see a CNC etched PCB from the various online resources, so obviously the finer
points that you've mentioned from first-hand are worth noting. Sounds like
not a great way to produce industrial prototypes after all, but still
good for hobbyists - the boards have *got* to look better than dead bug
prototypes !
Cheers,
Pete Restall
--- Original Message ---
I had access to a machine several years ago in college. I was never
able to get very good boards with it. Milling the PCB results in
grooves on the surface which makes a rough and unsightly board. There
is no solder mask or silkscreen. Unless the bit is brand new it will
do a poor job milling the copper, leaving burrs that will short out
traces. IIRC the minimum trace/space with this machine was 10mil/10mil
(0.01 inch) which precludes any modern parts. You can do two layer
boards but you can't do plated through vias or holes.
All-in-all it's a bum deal for a machine that is (at minimum) several
thousand dollars. It hasn't really caught on because people who have
the capital to pay for it are probably willing to pay to have boards
expedited, and are probably also doing "state-of-1970" circuit boards
that are too advanced for a milling machine to produce.
How long does it take to do a board that sort of size (roughly - obviously
it depends on track complexity) ? Is it comparable in speed to chemical
etching ?
I noticed you've milled just the gaps between tracks; have you tried to
take larger areas of copper off the board at any point - just etching the
gaps between tracks seems to be common on the CNC boards I've seen on the web,
but I personally don't like to leave floating copper. I'm just wondering if
it's too much hassle to do this (ie. takes too long) ?
How long do you reckon the tools will last before needing to be
replaced/sharpened ?
Cheers,
Pete Restall
--- Original Message ---
I built a CNC machine from a milling machine using EMC2 software and an
interface of my own design. While I did this orginially to do aluminum
with, I recently have been doing PCBs with it. For one board that uses a
2mm x 2mm chip with 3 pins on opposite sides I use a 20 degree engraving bit
for a gap of about .018". For less critical work I use a .029" end mill set
only deep enough to remove the copper and not the board. You can see a
board I am currently selling at http://www.dv-fansler.com/dvf%20Technologies/lx200_keypad_sim.htm along with
3 pictures. This board was done with the .029" end mill. I have done over
20 boards on the same end mill and they still look like the first one.
BTW - I do my layout in AutoCAD and then convert it to ngc with a free lisp
program called Realize.
On Sun, 28 Feb 2010 20:52:54 -0800, William "Chops" Westfield wrote:
> Here's a copy of the message I sent out several years ago...
>
>> I bought a used LPKF from eBay a couple years ago for hobbyist
>> usage. Less than half the price of the cheapest new unit...
>
> [snip]
It's looking more and more like it's going to be more hassle/time/money
than just sticking with chemical etching. But at least a CNC can be put
to other uses I suppose...
Hi Peter - the board takes about 20 minutes to mill out. You have to be
careful how you draw the board in AutoCAD, because it will mill the board in
the same order. I look for the longest continous line I can do first, then
the second, etc.
I have not tried to remove copper from a large area (except under an
antenna, receiver & transmitter I am working on for another project). How
much you can remove in a single pass depends on the diameter of the end
mill. If you wanted to take the time to program out large open areas to
remove with an larger diameter end mill and then progressively smaller areas
with a smaller end mill - it could be done - however I feel the time to draw
out the path's would be very time consuming.
I tend to run the end mill on that board a 2 inches/minute. The engraving
tool I run at .5 inches/minute.
The engraving tools tend to wear out quickly - so I use them only when I
need super fine pitch. The .029" end mill has produced about 20 boards and
shows no sign of wear at this point.
Kind regards,
David V. Fansler wrote:
> Hi Peter - the board takes about 20 minutes to mill out. You have to be
> careful how you draw the board in AutoCAD, because it will mill the board
> in
> the same order. I look for the longest continous line I can do first,
> then
> the second, etc.
It's faster than chemical etching for a board of this size (I would budget
~30 minutes for the entire process, from exposure to etched board). You
don't need as much room, and you don't need a darkroom environment.
On the other hand, chemical etching allows one to use any program that you
can print PCB files from, and the resolution is higher.
There are trade-offs I will freely admit. I use to etch boards back in the
1980's. I did not set up my milling machine for doing boards, but it does
them and so I thought I would use it.
On Mar 1, 2010, at 6:24 PM, David V. Fansler wrote:
> On the other hand, chemical etching allows one to use any program
> that you can print PCB files from
The LPKF software can start with a number of input formats, including
standard gerbers, and includes options to remove excess copper
(including using a combination of finer (more expensive) mills and
larger (cheaper) mills as needed.) This is the sort of thing I was
alluding to when I mentioned the differences between the commercial
PCB plotters and the hobbyist "my CNC mill can also make PCBs" variety.
One of the nice things is that you mount your blank material on the
machine, change tools occasionally, and when you're done you have an
etched, drilled, and cut-out PCB. There is a lot less "in between"
time than with chemical etching.
> One of the nice things is that you mount your blank material on the
machine, change tools occasionally, and when you're done you have an
etched, drilled, and cut-out PCB. There is a lot less "in between"
time than with chemical etching.
/>
I briefly looked at using the CD/DVD printing facility on many inkjet
pro\inters as a means of printing resist ink directly to PCB.
I found that various people has done whole pro\inter mods to allow
larger sized PCBs to be printed directly.
Using relatively little automation you could get much of the process automated.
Adding an NC drill using a dremel like device and an XY table would
give you the holes. I'm sure numerous designs are available already.
Speed can be adjusted to accomodate low power mechanicsms. Feedback
would allow accuracy if desired.
Cutouts etc not handled without extra effort.
Overall system could be cheapish and goodish and would probably prove popular.
For extra points add an automated through link placer with soldering.
The processes are well enough known and understood that I'm surprised
no one has come up with a machine that does the whole process "in one
hour" like film machines at drug stores.
1. Little 4x6 double sided PCB blanks in hopper (eurocard size)
2. Feed into a mini CNC machine for drilling
3. Dip into through hole plating tank
4. Spray with photo resist
5. Expose with laser, both sides
6. Develop
7. Etch
8. Clean
9. Back to CNC machine for final routing
It wouldn't have the finishing steps of mask or silkscreen (perhaps an
inkjet or laser engraver process can be added for some silkscreen -
maybe the laser station that exposes in #5 can laser etch too).
But I bet there'd be a (small) market for it, somewhat like the rapid
3D prototyping machines offices are buying.
And since a lot of stuff is machine assembled, there's not much need
for silkscreen and soldermask.
Feed that into a paste applicator/pick and place machine, then into a
3 zone oven, and you've got a 1 hour circuit shop.
>> One of the nice things is that you mount your blank material on the
> machine, change tools occasionally, and when you're done you have an
> etched, drilled, and cut-out PCB. There is a lot less "in between"
> time than with chemical etching.
> />
>
> I briefly looked at using the CD/DVD printing facility on many inkjet
> pro\inters as a means of printing resist ink directly to PCB.
>
> I found that various people has done whole pro\inter mods to allow
> larger sized PCBs to be printed directly.
>
> Using relatively little automation you could get much of the process automated.
>
> Adding an NC drill using a dremel like device and an XY table would
> give you the holes. I'm sure numerous designs are available already.
> Speed can be adjusted to accomodate low power mechanicsms. Feedback
> would allow accuracy if desired.
>
> Cutouts etc not handled without extra effort.
>
> Overall system could be cheapish and goodish and would probably prove popular.
>
> For extra points add an automated through link placer with soldering.
>
>
> Russell
>
philpem.me.uk
